Welcome to chapter fourteen. Every earlier chapter asked what the body needs. This one asks how you make something that meets that need, and the honest first answer is that the method should be the last decision, not the first. You already have a wide toolset: software and electronics, additive manufacturing, CNC and other subtractive work, moulding and casting, metalwork, glasswork and textiles. This chapter does not teach you to use them. It teaches you to choose between them, including the frequent case where the right answer is not to make anything.
A workshop shapes how its owner sees problems. Someone with a printer sees printable problems; someone with a lathe sees turned parts; someone who writes software sees an app. A technologist who is genuinely useful in a clinical setting has more than one of those tools and, more importantly, the judgement to pick the one the problem calls for. This chapter does three things. It gives you a decision framework that starts from the need. It walks through each technology family as a clinical setting sees it, including the traps that matter more in healthcare than in a workshop. And it explains what changes as an object moves from a one-off to a batch to a product.
Before any method is considered, write the need down and answer these questions; each answer removes options. How many, and how often? Tooling processes only pay at volume. What does it touch, and for how long? The contact category may remove the method entirely. How is it cleaned, by whom, with what? Porous materials, crevices and exposed electronics fail first here. What loads, and how does it fail? Printed and bonded parts are weakest across their joins, and many plastics and glass fail without warning. How accurate must it be? Who maintains it? Does it need power or logic? And which risk tier is it? A tier three object is not a fabrication question for you at all.
This diagram is deliberately pessimistic about making things, because that is how a clinical engineering department thinks. First: does a commercial product already meet the need? Then buy it, and finding it is a successful outcome. Would changing the process, training, layout or information remove the need? The lost item may be better solved by moving where it is stored. Can an existing item be modified? Sometimes the most useful contribution of all, but modifying a medical device can void its evidence, so that is clinical engineering's decision. Is the problem information, timing or coordination? Then software. A commercial product arrives with evidence, spares and a recall route. A made object arrives with you.
Now each technology family, as a clinical setting sees it. Figures are typical shop-floor orders of magnitude, not specifications.
Software is the cheapest thing to change and the hardest to see. Many requests that arrive as "can you make a" are really information problems. It is bad at anything needing force or physical protection, or reliability when the power or network fails, and its intended use can make it a medical device. Electronics let an object sense, decide and signal, and bring the most new obligations. Anything electrically connected to a person is IEC sixty-six oh one territory. Low-voltage, battery-powered, self-contained designs are the realistic space. Firmware is software, and electronics fail silently, so design what the user sees when a sensor drifts or a battery runs low.
In a clinical environment electronics face specific demands. Lithium cells store a lot of energy, so use protected cells, certified chargers and enclosures that tolerate swelling. Hospitals are dense with radio transmitters and sensitive monitors, so a device must neither disrupt nor be disrupted by them. Ingress ratings describe protection against water; wipe-cleaning in practice means fluid running into every seam. Alcohols, quaternary ammonium compounds and chlorine releasers crack some polycarbonate and acrylic housings, lift label adhesives and corrode contacts. Check the chemical, not the word wipeable. Realistic examples include switch-adapted toys, timers and prompts, training aids that light up when a step is done right, and test fixtures.
Additive manufacturing builds parts layer by layer, and it is one family among many. Filament printing is cheap and fast but weak between layers and porous. Resin printing gives fine detail and smooth surfaces, but washing and post-cure are critical. Powder-bed nylon makes strong parts without supports but traps powder. Metal powder bed makes titanium lattices on industrial equipment. Its genuine strengths are complex one-offs, geometry derived from imaging, internal channels, rapid iteration, and jigs for other processes. It is the wrong choice when a part must be cleaned repeatedly, when the load runs across layers, when a cheap standard item exists, or when the quantity justifies a tool.
Milling, turning, routing, drilling and profile cutting start from solid stock. Their great clinical strength is that the part is the certified stock material: fully dense, no internal voids, traceable to its melt, and finishable smooth enough to clean well. Milling and turning hold a few hundredths of a millimetre routinely. Laser cutting is very fast for sheet, but never laser-cut PVC, which releases chlorine compounds. Waterjet cuts metal and glass without heat. The traps: burrs and sharp edges injure skin and tear gloves, coolants must be removed, deep tapped holes trap soil, and some plastics craze when cleaned with alcohol after machining stresses.
These methods shape material against a tool that costs once and then produces many identical parts. Injection moulding makes thousands to millions of accurate parts at low unit cost, the dominant process for medical disposables, after tooling costing thousands to hundreds of thousands. Silicone casting makes soft, skin-like parts in small batches, but platinum-cure silicone is poisoned by sulphur, latex and some printed masters. Polyurethane casting needs care with isocyanates. Thermoforming makes shells, trays and splints, and orthotics workshops are essentially thermoforming shops. The most accessible members, silicone casting and vacuum forming, are also among the most medically useful, because simulators and grips need soft, cleanable surfaces.
Metalwork remains central to healthcare equipment: trolleys, frames, rails, brackets and instruments. 304 stainless is general purpose but pits in chloride environments, including chlorine disinfectants. 316L adds molybdenum for chloride resistance and low carbon for weld zones, and is the usual choice for wiped equipment. Titanium is light, strong and biocompatible but expensive. Aluminium is light and easy to machine, but alkaline washer detergents attack anodising, and it corrodes against stainless. Passivation, which removes free iron and restores the protective oxide, is routine after machining stainless, not optional. Continuous, ground welds clean well; stitch welds trap soil. Match fasteners to the base metal.
Glass is chemically inert, transparent, heat-resistant in its borosilicate form, supremely cleanable, and it breaks into sharps. Scientific glassblowers still make the one-off laboratory apparatus a printer cannot make in an inert, heat-proof material, and optics underpin endoscopes, microscopes and eye models. In patient areas, breakage risks mean glass is replaced by polymers unless transparency truly needs it. Textiles are the technology closest to soft tissue: straps, slings, covers and positioning aids. Every stitch hole is a path for fluid, so fluid-proof covers use welded or taped seams. Design for the actual laundry process. Hook-and-loop collects lint and soil, and every strap edge is a pressure source.
Wood, cork, leather and card are honest prototyping materials, excellent for concept models, jigs, and objects for home or community use where clinical cleaning does not apply. They are generally excluded from clinical areas because they are porous; where wood appears, in a wheelchair seat base for instance, it is enclosed and covered. Most good objects are hybrids: a machined frame with a moulded grip, a printed housing around a commercial sensor, a textile cover over a thermoformed shell. Each interface between methods is a crevice, a bond or a galvanic couple. Design for manufacture and assembly keeps hybrids sane: fewer parts, standard stock, features that make wrong assembly impossible, and every part traceable.
Now what changes as an object grows up.
A prototype's job is to answer a question; it can be crude, but it must be labelled and kept away from use. A one-off made for a specific person or setting, under the host's review, needs drawings, materials, cleaning instructions, a version and a named owner. A batch of thirty identical teaching models is no longer a craft object: every unit must match the reviewed design, so you need a written process, jigs, inspection criteria and records. A product placed on the market needs a manufacturer with a quality management system under ISO thirteen thousand four hundred and eighty-five, design controls, risk management and a regulatory route. That is a company's job.
Some processes produce output that cannot be fully checked afterwards: you cannot inspect every weld inside or test every moulding to failure. So the process itself is validated. Installation qualification asks whether the equipment is set up as specified; operational qualification, whether it produces acceptable output across its allowed settings; performance qualification, whether it does so consistently in normal production. Moulding, welding, sealing, bonding and sterilisation are typical candidates. And where material came from becomes part of the design: a mill certificate says what a batch is, and medical-grade polymers come with a commitment to tell you before the formulation changes. A hardware-store sheet comes with neither.
It is common, and correct, for a clinical engineering team to choose a commercial item over a better-fitting made one. The commercial item has a manufacturer responsible for it, evidence, validated cleaning, spares and a recall route. That is not a judgement on your skill; it is a judgement on continuity. The way to compete is to make continuity part of what you deliver: dimensioned drawings with native and neutral files, a bill of materials with grades and suppliers, process notes, programs and firmware versions, inspection criteria, use and cleaning instructions naming the agents tested, spares, end-of-life criteria, and an honest statement of what has and has not been tested.
Now the engineering bridge.
Every method has a fixed cost, tooling, programming and setup, and a variable cost per unit. Two methods break even where their totals are equal: the difference in fixed costs divided by the difference in unit costs. Take an invented cleanable clip. Printed and hand-finished, there is no tooling but each part costs about six pounds. Injection moulded, an aluminium tool costs about eight thousand pounds and each part about eighty pence. The break-even is about fifteen hundred parts. But a cost-only comparison is incomplete. The moulded part is dense, smooth, repeatable and certified, so for a repeatedly cleaned item the break-even on suitability may come long before the break-even on cost.
When several parts sit in a chain, their tolerances add. Four parts each plus or minus a tenth of a millimetre stack to four tenths in the worst case, or about two tenths statistically if errors are independent. The method's typical tolerance must be smaller than the chain allows, often the decisive reason to machine rather than print a part that mates with a commercial device. Surface roughness tells a similar story. Hygienic design commonly targets about nought point eight micrometres or smoother. Polished stainless and moulded parts are well under that; machined parts around it; resin prints a little above; as-built filament prints ten to twenty-five micrometres. A bacterium is about one micrometre, so some processes leave bacterium-sized valleys.
A common cylindrical lithium cell stores around nine to thirteen watt-hours, so a small four-cell pack holds a hundred and fifty to a hundred and eighty kilojoules in something phone-sized. Short-circuited, crushed, overcharged or defective, a cell can enter thermal runaway, which spreads and is very hard to extinguish. Use protected cells, a proper charger and mechanical protection, and no improvised packs on a ward or near oxygen. For a batch, the question is whether the next hundred parts will be in tolerance. Process capability compares the process spread with the tolerance; one point three three or more is a common target. Measure ten consecutive parts and you learn more than one good part could tell you.
Making happens in several places in healthcare. Clinical engineering workshops repair, modify and build one-offs under the hospital's procedures, and are the natural host for supervised work. Maker spaces are often the entry point. Orthotics and prosthetics labs are thermoforming, lamination and milling shops. Rehabilitation engineering combines electronics, metalwork and textiles, strongly led by users. Dental labs run one of the most mature digital manufacturing workflows in medicine. And sterile services decide whether your object survives reprocessing. Range is the asset. The person who can say thermoformed shell, machined hinge, cast grip and a small microcontroller program, or actually no object at all, is rare and valuable.
The educator now wants thirty elbow models, wiped between sessions with detergent and alcohol wipes, maintained by a technician. Printing would mean about a hundred and twenty parts with layer lines that collect soil, variable pivot fit, and labels that wear off. Laser-cut acrylic is fast and crisp but crazes and cracks with alcohol wipes at its stressed edges. Casting is unnecessary for rigid segments, and moulding tooling is far beyond break-even. The choice is CNC-routed acetal or HDPE segments with a stainless shoulder bolt pivot, a replaceable braided polyester cord, and labels engraved and colour-filled so the redundant coding survives cleaning. Make three first, wipe them a few hundred times with the real product, drop them, measure, and only then cut the rest.
The lesson is in two sentences. The method that made the best prototype was not the method that made the best batch. And the decisive fact, that the department's wipes would crack acrylic, came from asking how the object would be cleaned, not how it would be made. The handover then includes drawings and CAM files, a bill of materials, jig drawings, assembly and inspection instructions, cleaning instructions naming the agent tested, spares, end-of-life criteria, and a status statement: approved design, batch made on a stated date, three units wipe-tested, no clinical use.
For the exercises: run three needs, a feeding-pump holder for a wheelchair, a knee model for physiotherapy students and a care-home medication reminder, through the decision diagram, and let at least one end without fabrication. Compare four methods for a real object with the cleaning agent and quantity as explicit columns. Plot cost against quantity from real quotes. Wipe offcuts of five materials two hundred times with an alcohol wipe and inspect for crazing. Calculate process capability from ten of your own parts. And write a handover pack that someone else can use without asking you.
The limits. Making a batch does not make you a medical device manufacturer; placing a product on the market needs a quality system, a regulatory route and a legal entity that answers for it. A process you have run is not validated; validation is a documented, statistical activity. What a cleaning agent does to your material, in your geometry, is established by testing with the actual agent. Batteries store real energy, and anything connected to a person is out of scope. And orthotists, prosthetists, clinical engineers, dental technicians and sterile services staff hold the judgement for their fields; your methods are offered to them, not substituted for them. Next, measurement, imaging, and the diagnostic laboratory.